Skip to main content

Research Projects

The research projects of marinom GmbH are based on the extensive experience of our engineering teams in the development of autonomous maritime systems. Particular emphasis is placed on the experimental validation of novel approaches in control systems, sensor technologies and data analytics.
By combining technical expertise with experimental approaches, we develop solutions that enhance the capabilities of existing systems while establishing the foundation for future technological advancements.

AnFAng - Applied Research for Audio Forensic Investigations

Within the framework of the Audio Forensics (AuFo) program, funded by the Agency for Innovation in Cybersecurity GmbH, multiple complementary research approaches are being pursued, including the AnFAng project. The primary objective is to extract previously inaccessible metadata from low-quality audio recordings.
As part of AnFAng, a system is being developed for the near real-time analysis of audio signals. This enables investigators to derive not only the spoken content of recordings, but also additional acoustic indicators and metadata. For instance, it becomes possible to infer the likely location of a caller. Key application domains include emergency call analysis and security-relevant scenarios such as counter-terrorism.
Using robotic agents (RAFA - Robotic Agent for Field Acoustics), high-fidelity reference datasets are generated through precise measurements. By analysing signals prior to filtering, robust room-acoustic features can be identified. Even under adverse conditions, this approach enables the estimation of room geometry and material properties.

project content

Motivation

To date, research in audio analysis has predominantly focused on speech processing and speaker identification. In contrast, the systematic extraction and analysis of metadata from audio signals remains  largely unexplored. In particular, it is not yet well understood which characteristics of the recording environment can be inferred from such data. 
As the reconstruction of recording environments from low-quality audio has received limited attention, this project constitutes a high-risk research endeavour with substantial potential for methodological innovation in forensic science.

  • Project Objectives

  • Development of a system for near real-time analysis of audio signal
  • Extraction of room-acoustic features and metadata from low-quality recordings
  • Generation of a high-resolution reference dataset based on detailed spatial measurement
  • Support of law enforcement agencies in emergency response, counter-terrorism and security-critical scenarios
objective

Funded by the Agency for Innovation in Cybersecurity GmbH

financing

Designed by vitamin B² | Konzept- und Werbeagentur.